The Safe Motherhood Hospital program aims to improve maternal and child health and reduce mortality among mother and child. The program has been in place for 10 years, but there has been no impact evaluation of the program. In this research, we evaluate whether the program has an impact on maternal mortality and the cost of child delivery. The research uses Propensity Score Matching (PSM) technique that widely used for non-experimental studies. PSM has been a popular approach to estimate causal treatment effects, particularly when evaluating labour market policies and social programs. The technique is considered inexpensive and very useful for evaluating public projects in Thailand where projects have been implemented without a prior design for impact evaluation. From out analysis, we find that number of changes of director of the hospital, geographic factors, size of catchment area, socio-demographic structure of the district and hospital resources are statistically significant determinants of SMH program participation both before 2004 and 2005. Using various weighting and matching methods, we find that the SMH program has no impact on the proportion of maternal death to child delivery. The program, however, can reduce cost (i.e. length of stay for birth giving and labour cost) of participating hospitals. Given that the program does not allocate more resources to the participating hospitals, its guidelines can reduce the cost of child delivery without increasing the death rate. This could be considered as a success of the program. The findings also show that the program encourages referral. However, referral among community hospitals may be contaminated by the way the NHSO financed referral cases.